Declarative Debugging of Wrong and Missing Answers for SQL Views

Abstract

Abstract. This paper presents a debugging technique for diagnosing errors in SQL views. The debugger allows the user to specify the error type, indicating if there is either a missing answer (a tuple was expected but it is not in the result) or a wrong answer (the result contains an unexpected tuple). This information is employed for slicing the associated queries, keeping only those parts that might be the cause of the error. The validity of the results produced by sliced queries is easier to determine, thus facilitating the location of the error. Although based on the ideas of declarative debugging, the proposed technique does not use computation trees explicitly. Instead, the logical relations among the nodes of the trees are represented by logical clauses that also contain the information extracted from the specific questions provided by the user. The atoms in the body of the clauses correspond to questions that the user must answer in order to detect an incorrect relation. The resulting logic program is executed by selecting at each step the unsolved atom that yields the simplest question, repeating the process until an erroneous relation is detected. Soundness and completeness results are provided. The theoretical ideas have been implemented in a working prototype included in the Datalog system DES.

Optimization Techniques for Algorithmic Debugging

“…In Algorithm 12, the new code in the sqrt method includes a call (line 6) to the recursive method sqrt loop that implements the loop (lines [9][10][11][12][13][14]. This new recursive method contains the body of the original loop (line 10).…”

“…For further information about bugs of type 2 (missing results), which are more related with logic programming and non-deterministic systems, we refer the interested reader to specific works such as [89,11,87]), which will provide more detailed explanations and examples about how to treat them.…”

“…Starting Debugging Session... Fortunately, using our balancing algorithm, the debugger can transform the EF in Figure 5.8 to produce the EF of Figure 5.9, where Nodes (5) and (6); (8), (9) and (10); and, (11) and (12), have been collapsed. The tool automatically detected that they form chains and collapsed them to form a single question.…”

“…int numChildren = this.getNodeChildren(); (9) (10) // Heaviest child (11) int indexChild = -1; (12) double weightChild = -1; (13) for (int index = 0; index < numChildren; index++) { (14) this.moveNodeToChild(index); (15) double weightCandidate = this.node.getWeight(); (16) if ( (2) return true; (3) if (weightNode1 > weightNode2) (4) return rootWeight ≥ weightNode1 + weightNode2 -individualWeight; (5) return rootWeight < weightNode1 + weightNode2 -individualWeight; end -Trusted if the method associated with this node is trusted [65] and, therefore, cannot contain the bug, and -Unknown if this node has been marked as unknown, e.g., because its associated question was too difficult.…”

“…In the panel we see that the initial values were [11,15,10,18,7], and the final values are [7,10,15,11,18]. Therefore, the user should mark this computation as wrong.…”